Previous studies have firmly established no NMR spectroscopy as a useful method for studying intramolecular hydrogen bonding. However, assessment of the influence of intra
molecular hydrogen bonding on structure and the resulting 170 chemical shift changes have not been fully elucidated. We have completed a study on a series of 2'- and 4' -amino and amido substituted acetophenones (49 to 54) and a-amido substituted acetophenones (55 and 56) which shows that the no NMR data must be interpreted as a combination of torsion angle changes, electronic considerations and hydrogen-bonding shielding effect. 18
&H H N.._ I R,
49,51,53 50,52,54 55,56
The 170 NMR chemical shift data for 2'- and 4' -aminoacetophenone (49 and 50) 2'
and 4' -acetamidoacetophenone (51 and 52), 2'- and 4' -trifluoroacetamidoacetophenone (53 and 54), and a-acetamido- and a-trifluoroamidoacetophenone (55 and 56), obtained at natural abundance (0.5 M) in acetonitrile at 75°C, are summarized in Table 4. The ketone signals appeared between 519 and 551 ppm and those for the amido carbonyls appeared between 320 and 373 ppm as expected. For the para substituted compounds, the ketone and the
.... TABLE 4 2 170 Chemical Shift Data18
( ± 1 ppm) for Compounds 49 to 56 in Acetonitrilea at 75° (and in CHCI)b 22 Compound Compound Compound ~ R, fix no. fiy no. fiR, fiRz no. R, fiz fiRRz 3 ~ 524.1 518.7 49 H 50 H :::ti (518.0)b (513.0)" ~ 539.5 51 COCH535.4 373.3 52 COCH366.3 55 COCH532.6 337.1 5 3 3 ~ 53 COCF531.0 343.3 54 COCF551.0 339.1 3 3 56 COCF530.0 319.6 3 (528.9)b (341.5)b (542)b (338)bã' ~ .g '-<::! • Referenced to external water at 75°, 2-butanone as internal standard (558 ppm). sã b Referenced to external water at 35°; 2-butanone as internal standard (554 ppm). 0 ' Saturated solution. ~ l:l ;::: <=;ã [ <:;ã ~
amido carbonyl chemical shifts were as predicted based upon electronic effects reported in each series. However, the no NMR chemical shifts for the ketone signals of the ortho isomers 51 and 53 were shielded relative to those of their para isomers in contrast to that of 49 relative to 50. Since the electronic effect for the para substituent of compound 56 is essentially zero, the significant shielding (20 ppm) noted for the ortho isomer 53 is clearly the result of intramolecular hydrogen bonding. For the N-acetyl compounds, the ketone signal for the ortho compound 51 appears only slightly shielded (5 ppm) relative to 52 which could be interpreted to indicate a weak hydrogen bond. In contrast, the ortho amino compound 49 is deshielded relative to its para isomer 50, misleadingly indicating a lack of hydrogen bonding (vide infra). The data for 49 and 50 were qualitatively in agreement with those in dioxane reported by Fiat. 14 To check for intermolecular hydrogen-bonding effects of solvent, the no NMR spectra of the amino compounds 49 and 50 and the trifluoroacetoamido compounds 53 and 54 were taken in dichloromethane (Table 4). The net differences were essentially the same in both solvents. Compounds 55 and 56 were chosen to test for the effect of five-membered ring intramolecular hydrogen bonding on 170 NMR data. The ketone chemical shifts for 55 and 56 were essentially identical and appear shielded relative to analogs indicative of hydrogen-bonding effects.
Molecular mechanics (MM2) calculations which include hydrogen bonding considera
tions were carried out on compounds 49 to 56. The molecular mechanics calculations on these systems predict that the ortho compounds have significantly different geometries than their para isomers (Table 5). The calculated torsion angles for the functional groups for the para isomers 50, 52, and 54 are essentially identical to those reported for the respective parent compounds.6 ã25 Calculations for the ortho isomers 49, 51, and 53, which neglect hydrogen bonding' show larger torsion angle changes (01 c2c3c4 == 35-31 °) than those shown in Table 5, analogous to those reported for ortho alkyl groups. 6 •25
TABLE 5
MM2-Calculated18 Hydrogen-Bonding Distances, Bond, and Torsion Angles for 49, 51, 53
Compound
no. R H. ... O,(A) O,H.Ns o,c2c3c. H.N.c.c3 O,H6N5C4
49 H 1.97 122.3° -12.8° 1.20 -16.2°
51 COCH3 1.95 120.1° -16S 19.9° -52.7°
53 COCF3 1.97 117.3° -12.1° 24.7° -53.7°
Interestingly, the hydrogen-bonding distances are predicted to be essentially constant for 49, 51, and 53. The calculated bond angles for these six-ring, hydrogen-bonded systems are roughly 120°. Hydrogen bonding is predicted to reduce the torsion angle for the ketone carbonyls to between 12° to 17° from that expected based only on steric interactions. We have shown that torsion angle rotation of aryl ketones causes significant deshielding for the carbonyl signal (0.84 ppm/deg.)6 It should be noted that earlier investigations had concluded that electronic contributions at the ortho position are substantially less than at the para position; this was due in part to model studies using o-methoxy groups, 14•15b and the as
sumption of no conformational changes. At the time of those studies the importance of torsion angle changes was not appreciated. The result of torsion angle changes for the ortho
106 170 NMR Spectroscopy in Organic Chemistry
analogs is to partially mask the electronic contribution because the effect on 170 NMR chemical shift is in the opposite direction. Since for aryl carbonyls torsional rotation yields deshielding effects, while hydrogen bonding effects are shielding, the chemical shifts ob
served are a combination of these competing influences. The contribution of hydrogen bonding to the chemical shift can be estimated by subtracting the value for the ortho isomer from the sum of values of the torsion angle-rotation contribution and the chemical shift data for the para isomer (use of para isomer value corrects for electronic substituent effects).
This approach yields shielding effects of 5, 18, and 30 ppm, respectively, for hydrogen bonding in 49, 51, and 53. This trend parallels the relative acidity ofthe N-Hs. The shielding value for compound 53 begins to approach that noted for phenolic analogs. 14 Casual in
spection of the data for 49 and 50 might suggest that hydrogen bonding is unimportant in 49; however, this method of analysis shows that hydrogen bonding contributes significantly to the structure of 49. This approach assumes that the electronic effects of ortho and para substituents are identical; however, as a result of torsion angle differences, the electronic contribution for the ortho isomer may be somewhat reduced. If valid, this will lead to even larger shielding contributions for hydrogen bonding in these systems.
Molecular mechanics calculations for 55 and 56 showed identical hydrogen bond dis
tances and angles (OHN) of 2.18 A, and 105°. The amides in this case were predicted to be essentially planar unlike the aryl acetamides which were predicted to be slightly pyramidal.
These results may explain, in part, the shielding seen for the amido carbonyl signals of 55 and 56 compared to that of the acetanilides. 25
The apparent lack of competition of intermolecular hydrogen bonding to the solvent (acetonitrile) is intriguing. It is doubtful that the solvent is inert; more likely the 2' and 4' isomers are experiencing analogous interactions to yield the apparent insensitivity. Intra
molecular hydrogen bonding of an amide NH to a methoxy group was deduced from a small solvent dependency; however, torsion angle dependence was not considered. 17 The 170 NMR data seems sensitive to hydrogen bonding even in these systems which are conformationally mobile and appear to experience considerable changes in geometry. The data clearly show that torsion angles must be considered when analyzing hydrogen-bonding phenomena by
170 NMR spectroscopy. Despite the mobility of this system, the170 NMR data indicate that larger shielding effects are seen with the more acidic NHs.
An 170 NMR study of intramolecular hydrogen bonding of N-H to carbonyls in sub
stituted fluorenones and anthraquinones26 (Scheme 10, Tables 6 and 7) supports our previous conclusions. 18 •27 In these cases torsion angle considerations have been minimized; the in
tramolecular hydrogen bond distances and angles would be expected to vary between the fluorenone and anthraquinone systems. Applying the previously described method of analysis
TABLE 6
170 NMR Data26 of Amino and Amido Substituted Fluorenones in CH3CN at 75oca
o(C=O) ii(NHCO)
No. Compound ppm ppm
42 Auorenone 517.3
(510.1)'
57 1-NH,-Auorenone 486.3
58 3-NH,-Auorenone 483.3
59 1-NHCOCH3-Auorenone 500.8 378.1
60 3-NHCOCH3-Auorenone 507.1 371.0
61 1-NHCOCFrAuorenone 499.5 348.7
' Reference 15b in CDCI3 at 40°C.
R = H, Rc, COCF3
SCHEME 10
TABLE 7
170 NMR Data26 of Enriched Amino and Amido Substituted Anthraquinones in
CH3CN and 75°C8
3(C=O)
No. Compound ppm
62 Anthraquinone 524
(53l)b
63 l-NH2-Anthraquinone 485.5
515
64 1-NHAc-Anthraquinone 496.5
525.5 65 l-NHCOCF3-Anthraquinone 494
532
• The quinone carbonyls were enriched; hence, the -NHCO carbonyl signals for 64 and 65 were not detected.
b Reference 15a from toluene at 95°.
to the N-acetyl and N-trifluoromethyl fluorenones 59 and 61, the upfield shifts of 4 and 17 ppm, respectively, are attributable to hydrogen bonding. These values are much less than deduced for the analogous acyclic examples (18 and 30 ppm). This result is consistent with MM2 calculations that predict the intramolecular hydrogen bonding distances are longer for the fluorenone cases (Table 8) than for those of the acetophenone and anthraquinone cases.
TABLE 8
MM2-Calculated26 Hydrogen-Bonding Distances, Bond and Torsion Angles for Fluorenones (57, 59, 61) and Anthraquinones (63 to 65)
1
0 0
5 H 6 H COCH3
1.92 1.91
121.6°
118.7°
120.7 120.9
oo
0.9°
oo
27.1°
oo
-43.6°
N" ... COCF3 1.93 115.8° 120.8 0.9° 30.7° -48.0°
R
1 H 2.21 122.4° 127.7 o• oo
0 COCH3 2.02 132.5" 127.6 4.5° -8.8°
6 COCF3 2.07 127.1° 127.4 18.1° -30.9°
s,H
N ' R
108 170 NMR Spectroscopy in Organic Chemistry
Interestingly, this method of analysis predicts essentially no contribution of hydrogen bonding to the 170 NMR chemical shift of the carbonyl of 1-aminofluorenone (57). The MM2 calculated hydrogen bond distances (Table 8) are substantially longer than those for the acetophenone systems which are consistent with the observed 170 NMR results.
The anthraquinone data,26 on the other hand, clearly show intramolecular hydrogen bonding for both the amino and amido cases. Analysis as discussed above yields intramo
lecular hydrogen-bonding shielding effects of 6, 15, and 29 ppm, respectively, for 63, 64, and 65. The 170 NMR shielding effects attributable to hydrogen bonding are essentially identical with those of their acetophenone analogs, and the MM2 predicted hydrogen-bonded structures show similar geometric parameters (Table 8).
Our intramolecular hydrogen-bonding studies of amino and amido groups to the carbonyl moiety clearly show that the factors which determine 170 chemical shift can be deduced.
The contribution of hydrogen bonding to chemical shift can be quantitated only after cor
rections for substituent effects (electronic effects) and torsional variation have been per
formed. Hydrogen-bond induced chemical shifts (A3) correlate reasonably well with the acidity of the N-H and seem to show a dependence on the hydrogen-bond distance. The angular dependence of the hydrogen bond on the 170 NMR data in systems such as these awaits further experimentation.
Applying the above approach to intramolecular hydrogen-bonded phenols (to carbonyls) allows the quantitation of the influence of hydrogen bonding on 170 NMR chemical shifts therein. The 170 NMR data for a series of ortho-hydroxy acetophenones in acetonitrile are listed in Table 9. 27 Interestingly, the advent of high resolution instruments allows the detection of proton-to-oxygen coupling in the phenol signal. For example, Figure 1 shows proton coupled and decoupled spectra of 2-hydroxyacetophenone. Coupling is expected under con
ditions where fast intramolecular proton transfer between the two oxygens occurs without proton spin exchange. 13 Other solvents, lower temperatures, and/or large broadening factors can obscure the coupling. The only previous example of similar coupling was deduced for the enol of 2-acetylcyclohexanone by line-shape analysis, and a J value of 77Hz was noted by Gorodetsky. 13 In the present examples the observed coupling constants ranged from 75 to 90 Hz (Table 9).
Correcting the 170 NMR carbonyl chemical shifts for torsion angle rotation and electronic effects as described above for amino and amido systems yields the shielding contributions for hydrogen bonding (Ao) shown in Table 9. For example, the hydrogen-bonding contri
bution to the 170 carbonyl chemical shift of 67 is estimated as follows: the parent system (acetophenone) chemical shift is 552, the electronic effect of a para methoxy group is - 16 ppm, the electronic effect of the ortho-hydroxy group is taken as -21 ppm, and the torsion angle contribution is + 3 ppm, which gives a predicted chemical shift of 518 ppm for 67, assuming no effect of intramolecular hydrogen bonding. The observed chemical shift of 466 ppm is 52 ppm shielded from the calculated value. For all the systems, the range in con
tribution to chemical shift from hydrogen bonding is 43 to 67 ppm with the average effect near 55 ppm. Presumably, the range of Ao is in part due to the error in approximations of torsion angle and to changes in the acidity of the phenol as a consequence of substituent.
Interestingly, the predicted 170 NMR hydrogen-bonding contribution for 2,2' -dihydroxy benzophenone 46 is 53 ppm which is consistent with the formation of only one hydrogen bond to the carbonyl group in acetonitrile. However, these Ao values are less accurate as a result of the difficulty in estimation of the torsion angle changes in the benzophenones by molecular modeling methods. Previously, the 170 chemical shift data in chloroform for 46 were interpreted to show two hydrogen bonds to the carbonyl group. 15b The data in chloroform
TABLE 9
170 NMR Data27 ( ± 1 ppm) for Intramolecular Hydrogen-Bonded Phenols and Related Compounds in CH3CN at 75oca
No. Compound 8(C=0) 8(0H) J(OH) Ji8b
16 o-OH-Acetophenone• 491 85.5< 86 43
17 p-OH-Acetophenone' 531 88
18 2, 4-Dihydroxyacetophenone' 463 90, 94.5'ãd 50
19 2,6-Dihydroxyacetophenone• 481 91 51
66 2-0H ,6-0 Me-Acetophenone 484 89'; [62.5 (OMe)) 87 58
67 2-0H ,4-0Me-Acetophenone 466 91'; [67 (OMe)] 75 52
44 Benzophenone 552
45 2-0H-Benzophenone' 492 85 51
(486)' (84)' (36)'
46 2 ,2' -Dihydroxybenzophenone' 475 82 53
(444)' (80.5)' (57)'
68 2-Acetyl-l-Naphthol 466 90' 80 67
69 1-Acetyl-2-Naphthol 512 93 60
70 Propiophenone 540
71 2-0H-Propiophenone 481 85' 91 45
• See Tables 3 and 4 for values in different solvents.
b Calculated 170 NMR hydrogen bonding shielding effect; see text.
' Splitting of OH signal.
d Coupling was apparent; however, signal overlap prevented quantitation of J.
' In CDC13 ; see also Table 3.
are reproducible and substantially different from those obtained in acetonitrile. The ,:l() values in this system suggest the formation of two hydrogen bonds in chloroform for 46. Clearly, the 170 NMR hydrogen-bonding-induced shift data (94 ppm) for the 9-carbonyl signal of
I ,8-dihydroxyanthraquinone 72 in toluene at 75o are consistent with the formation of two intramolecular hydrogen bonds to the carbonyl group. 28 Also, this result is consistent with the X-ray structure of 1 ,8-dihydroxyanthraquinone. 29
444 80.5 (CDCL3) 395 91.5 (toluene) 475 81.7 (CH3CN)
vv ~
* 532 0
46 72
Application of corrections for substituent effects (electronic effects) and torsion angle rotation allows the determination of the contribution of 170 NMR chemical shifts of carbonyl groups arising from intramolecular hydrogen bonding. Analysis of the geometry of the hydrogen-bonding array must also be performed since less than optimum geometry will also
110 170 NMR Spectroscopy in Organic Chemistry
I I I I
200 I I I I I I
100
I I I I I I
0 PPM FIGURE I. Coupled and decoupled spectra of o
hydroxyacetophenone in acetonitrile at 75°C. 21
dramatically reduce the effect. Large upfield shifts (ca. 55 ppm) are attributable to intra
molecular hydrogen bond formation in acidic systems (phenols). Intramolecular hydrogen bonding involving the amido NH of trifluoroacetyl derivatives causes 30 ppm upfield shifts which are approximately twice as large an effect observed for acetyl NH's participation in intramolecular hydrogen bonding (18 ppm). The contribution to chemical shift from amino NH intramolecular hydrogen bonding is much smaller (5 ppm). Thus the magnitude of the observed shifts parallel the acidity of the hydrogen bond donor. Clearly, 170 NMR meth
odology is sensitive to intramolecular hydrogen bonding to carbonyl groups and provides a promising approach for future studies.